CV

Curriculum vitae. Use the button above to download the PDF version.

Contact Information

Name Ruixin Song
Professional Title Product Engineer & Founding Team Member
Email ruixinsong21@gmail.com

Professional Summary

Working on representation learning, graph learning and physics-informed AI, with a background in spatiotemporal trajectory modelling for maritime data.

Experience

  • 2026 -

    Vancouver, BC

    Product Engineer & Founding Team Member
    GradientX Technology, FinTorch
    • Co-founded FinTorch, an AI-powered financial copilot exploring personalized financial guidance, recommendation systems, and decision-support agents for individual users.
    • Designed and implemented the core technical architecture, including a RAG- and LangChain-based multi-agent pipeline for conversational financial question answering, personalized recommendation, and context-aware decision support.
    • Collaborated on research problem formulation, technical roadmap planning, product validation, and iterative evaluation within a cross-functional founding team.
  • 2024 - 2026

    Halifax, NS

    Research Assistant (contract full-time)
    Dalhousie University, AISViz, MAPS Lab
    • Conducted spatiotemporal trajectory representation learning for similarity computation.
    • Built scalable ETL pipelines for 15 years of shipping data in PostgreSQL/TimescaleDB, cutting storage by 34% with advanced indexing and partitioning.
    • Refactored Rust modules in AISdb, improving performance and correctness; maintained CI/CD workflows with GitHub Actions.
    • Supported lab research and collaborated with government and academia to deliver high-resolution maritime datasets for regulatory and scientific use.

Education

  • 2021 - 2024

    St. John's, NL

    M.Sc. (thesis-based)
    Memorial University of Newfoundland
    Computer Science
    • Computer Graphics, Machine Learning, Research Methods
    • Thesis: Temporal Analysis and Gravity-Informed Marine Traffic Forecasting for Non-Indigenous Species Risk Assessment Through Ballast Water
    • Advisor: Dr. Amilcar Soares Junior
  • 2020 - 2020

    Montreal, QC

    Graduate Diploma program
    Concordia University
    Computer Science
    • Computer Architecture, Algorithms, Academic Writing
    • Transferred to Memorial University after the first semester.
  • 2016 - 2020

    Shanghai, China

    B.Eng.
    Shanghai Ocean University
    Spatial Information and Digital Technology
    • Thesis: Privacy-Preserving Electronic Voting System Using Homomorphic Encryption
    • Advisor: Dr. Lifei Wei

Awards

  • 2024
    Fellow of the School of Graduate Studies
    Memorial University of Newfoundland
  • 2021
    Graduate Fellowship, $16,000/yr
    Memorial University of Newfoundland

    Held 2021–2023.

  • 2022
    Best Poster Award (1st place) in Computer Science
    Scientific Endeavours in Academia Conference
  • 2017
    People's Scholarship
    Shanghai Ocean University

    Held 2017–2019.

Projects

  • TransformerGravity

    A gravity-informed deep learning framework with self-attention for global marine traffic forecasting. Funded by NSERC and Memorial University.

    • Designed and implemented a deep learning framework in PyTorch combining stacked Transformer architecture with graph-based representations to forecast global shipping traffic patterns.
    • Improved prediction accuracy by 13% over deep-learning baselines and 50% over traditional machine-learning models; optimized training code to improve performance stability.
    • Built a data preprocessing pipeline (NumPy, Pandas, SciPy) supporting analysis of over 3.8 million records, and trained comparison models using scikit-learn.
    • Led project implementation and technical documentation; co-authored a paper in Scientific Reports and delivered a spotlight talk at the 36th Canadian Conference on Artificial Intelligence (2023).
  • BWRA for Bio-Invasions

    Improved the ballast water risk assessment (BWRA) model used by Transport Canada. Funded by Fisheries and Oceans Canada, Transport Canada, and Memorial University.

    • Spatially analyzed multilayer sea surface environmental data and matched it to over 8,300 global ports.
    • Refined Transport Canada’s BWRA model using fine-grained environmental data, demonstrating statistically significant improvement (Wilcoxon signed-rank test, effect size > 0.5).
    • Reduced model runtime by over 80% through code optimization, improving feasibility for operational use.
    • Collaborated on an interdisciplinary team combining technical and biological expertise, resulting in a co-authored paper in Biological Invasions (Springer Nature, 2023).

Teaching

Activities

Skills

Programming languages: Python, Rust, Julia, R, C++, JavaScript
Tools: scikit-learn, PyTorch, SciPy, WebGL, D3.js, graph-tool, NetworkX, PostgreSQL

Languages

Mandarin : Native
English : Professional working proficiency (C1)
French : Classroom study (A1–A2)

Interests

Research interests: Representation Learning, Graph Learning, Physics AI